Abstract
Introduction and Argument
During the first half of 2008, two of Africa's shining lights witnessed mass violence. In Kenya, the demon of tribalism, which many hoped had been exorcised at the end of Moi's rule, resurfaced, threatening to derail the past five year's economic and political achievements. Although xenophobic violence was not new to South Africa (Crush 2008; Harris 2001), the May 2008 attacks revealed the government's fundamental inability to protect the security and welfare of all its residents. In both countries, the police's inability or unwillingness to stem the violence raised the question of “who controls the streets?”.
Answering this question means addressing what contemporary ethnic and xenophobic violence says about the nature of African society and states, as well as the security of those ostensibly depending on them for protection. Our cursory comparison of Kenya and South Africa can only be a tentative response. In it we find remarkable similarities rooted in colonially imposed ideas of territory and its relation to political and economic privilege. These have allowed discourses of indigeneity to be mobilised to exclude competitors from national or sub-national economic and political resources. In such environments, there are irreconcilable conflicts between conceptions of national or universal rights (economic, social, and political) and beliefs that such rights are inextricably tied to someone's territorial origins and physical location. Where such understandings are linked to sub-national space, ethnic conflict ensues. Where they accept, naturalise and reify national boundaries, the result will be xenophobia. In both cases, such perspectives on rights lead to varying degrees of social and institutional exclusion. As the displacement of hundreds of thousands in Kenya and South Africa shows us, Fanonian violence is never far from such campaigns.
As similar as the two countries are, the 2008 violence also reveals critical differences in the structure and object of the violence. In both cases, the attacks reflect a mix of criminal opportunism and political intention. However, the Kenyan case suggests a more coordinated and state-centred campaign. There the violence was more or less controlled by Raila Odinga's Orange Democratic Movement as a way to contest what was widely presumed to be a stolen presidential election (International Crisis Group 2008). Where the courts are slow and lack autonomy, such violence represents a logical – if unfortunate – means of protest and political influence (see Kimenyi and Ndung'u 2005: 146). In this case, the objective was control over the central state. When this was at least partially achieved, the violence all but stopped.
Although the South African violence produced fewer deaths and less displacement, it represents a more insidious form of political mobilisation. Unlike Kenya's violence, the South African attacks reflect a territorialised and anti-state political culture. The goal was not control over the central state nor was the violence coordinated by anyone other than local gangsters and politicians. Even as the attackers evoked the nation in expelling outsiders from their communities, the violence was not about establishing a national political culture or protecting the nation-state. Rather, it helped to reinforce a territorialised heterodoxy composing multiple systems of rights and systems of rule each attempting to make exclusive claim to territory and the resources held within. At times this meant claiming resources provided by the central state – particularly houses or local offices – but not the central state itself. Indeed, the success of the violence will ultimately reduce the central state's ability to integrate and regulate the territory it ostensibly controls.
Our Approach, Our Data
A short article can not reflect the depth and horrors of the violence in either Kenya or South Africa. Nor can it provide a detailed, comparative explanation for 2008's stinging events. Instead, we attempt to make sense of the violence in ways that point to broader conclusions about the nature of violence and society. In doing so, we blend a micro-level analysis of people's spatialised subjectivities with broader insights into institutional structures and regimes of control and regulation (cf. Allen 1997; Agnew 2005; Davis 1998).
Our work draws on a mix of secondary and primary sources of data. To our discredit, the Kenyan component is informed largely by news reports and accounts provided by non-governmental organisations and first hand witnesses based in South Africa. We complement these with our personal experiences of living and working in East Africa. The South African component is considerably more robust. This article specifically draws on four months of fieldwork in Gauteng Province and the Western Cape. To understand the triggers of the violence, we selected five sites where xenophobic violence occurred between January 2007 and June 2008, and two sites where the presence of foreign nationals has not led to significant violence. At each site, we interviewed South African residents, foreign nationals who reside or resided in the same locations, relevant government officials, community leaders, and representatives of different civil society organisations operating in the selected areas. In addition to in-depth interviews, we held two focus groups (of five to ten members) in each of the communities. In total, close to 300 people (including young adults, women and the elderly) participated in the study. As a result of our expertise and the availability of data, the remainder of the article is decidedly, if regrettably, biased towards South African events.
Historical Precedents: Violence, Segregation, and the Politicisation of Space
The 2008 violence in South Africa and Kenya cannot be understood separated from extended histories of racial and ethnic discrimination, predatory politics, and violent resistance. In Kenya, the colonial appropriation of the productive (and comfortably cool) highlands resulted in the dislocation of tens if not hundreds of thousands of people. Some ended up in labour or native reserves; others were spread across the country (Kennedy 1992). Further South, apartheid and the Group Areas Act separately attempted to create ethnically and racially pure “Bantustans” within South Africa's sovereign territory. In both countries, parts of the cities were all but off limits to the countries' “native” population while inter-ethnic (let alone inter-racial) mixing was discouraged or overtly prohibited.
Much has been written about how colonial racial and ethnic divides have translated into post-colonial conflict (see, for example, Mamdani 2002). Although such divisions do not translate automatically to violence, naturalised divisions remain an all too frequently mobilised resource. Much less attention has been paid to how fragmentary identities became spatially rooted and how rights to space become resources for ethnic or racial mobilisation. An article of this length can not fully address the issue of converting land into socialised, politicised space (cf. Lefevbvre 1991). Rather, it simply roots the processes in both countries' settler histories, which are replete with extensive systems of domestic exclusion and territorial demarcation. In Kenya this was accomplished through the creation of labour reserves and other forms of prohibitions (see Murungi 1995). In South Africa, it was generated by grand apartheid's dangerous folly. These systems of divide and rule served as antecedents to contemporary socio-political configurations and post-independent approaches to outsiders, whether from another part of the country or across a border.
As before, citizens and officials in both countries continue to see unregulated human mobility as a threat to the citizenry's economic and physical well-being; an individual's immutable geographic or cultural point of origin continues to determine insider or outsider status. Such attitudes are further reinforced by urban planners' Malthusian fears that social mixing and uncontrolled movement is both socially and politically unsettling. That rights to land – whether agricultural or urban – continue to be so highly politicised further reinforces the value of territory as a point of political contestation.
Thrown into the mix of ethnic and territorial divides is an extended history of violent politics in both countries and the inability – or unwillingness – of the central state to fully monopolise the legitimate means of coercion. Whether it is Nairobi's Mungiki or a long history of politically motivated violent attacks mobilised through claims to land and ethnic purity, violent struggles over land have been regular features of Kenyan political life (see Human Rights Watch 1993). The International Crisis Group's (2008: 11) report reflects this, speaking of, “deeply entrenched, long-festering anti-Kikuyu sentiment” that has often resulted in violent conflict between “natives” and outsiders'. Kanyinga (2000) similarly argues that there has been a long history of violence linked to an elite, ethnic “other” making land claims outside of their “traditional” or “ancestral” homelands. Although Kimenyi and Ndung'u (2005: 125) argue that “conflicts are a recent phenomenon in Kenya”, such an assertion is hard to sustain. The country was forged from violence and a certain level of violent suppression and expression has always been part of the country's politics.
Although South Africa is justifiably famous for its peaceful transition from Apartheid, the transition was preceded by an extended period of violence founded on both ethnicity and rights to space. During the Apartheid era, the threat of violence – whether “vertical” (state against citizens) or “horizontal” (citizens or rival factions against each other) – saturated the lives of South Africans residing in the volatile, tightly policed townships (Hamber 1999). In the wake of the ANC's unbanning, vertical violence was largely overshadowed by horizontal forms enacted largely through armed conflicts between supporters of the African National Congress (ANC) and the Inkatha Freedom Party (IFP) (Hamber 1999).
The end of the political struggle did not represent an end to horizontal violence as individuals and groups continued exploiting coercive means for personal and non-political ends, often under the cover of crime-fighting. Violence in South Africa's townships should be understood against this background and its links to years of social and economic disadvantage, repressive policing, criminal predation and a consequent recourse to vigilantism cum mobsterism (Kynoch 2005). Indeed, the countries' respective economic capitals – Nairobi and Johannesburg – have global reputations for lawlessness and gangsterism (Lacey 2001; Anderson 2002; Steinberg 2004).
In both countries, the state has been either complicit in past violence or has been unable to stop it, something that has further normalised violence as a political tool and fostered a culture of impunity (see, for example, Human Rights Watch 1993). In its explanation of the past, the Waki report (CIPEV 2008: 2) speaks of a “culture of impunity and a consistent escalation of violence.” In South Africa, the National Prosecuting Authority has only taken forward a handful of cases of apartheid-era violence. Across the country, more quotidian – if only slightly less brutal – cases are frequently dropped or not investigated at all (Altbeker 2007). In South Africa, the culture of impunity is particularly pronounced regarding xenophobic violence. Non-nationals have been repeatedly attacked in South Africa since 1994 but few perpetrators have been charged and fewer convicted. In some instances, state agents have actively protected those accused of anti-foreigner violence. In Masiphumelele, a township near Cape Town, the former Western Cape Provincial Premier, the Member of the Executive Committee and the local Police Commander intervened to secure the release of businesses owners who had been arrested after xenophobic violence in 2006. Similarly, before, during and after the May 2008 violence, some arrests were made at the different scenes of violence but most of them were released without charge due to community protests and mobilisation. The actual and perceived impunity with which perpetrators of xenophobic violence are seen to act can only continue to encourage the ill intentioned to attack foreigners.
In both countries, new regimes have either not been able to shake off such “heavy” institutional legacies and still use the same “technologies of alienage”, or have been authors of new forms of marginalisation that are reinforcing a localised/territorialised, nationalistic and ethnic understanding of rights. Such an understanding is in turn leading to a violent “nativist revivalism” (see Mbembe 2006).
Understanding Violence Over, Beyond, and against the State
While there are strong historical similarities between the two countries, these alone do not explain the appearance or nature of the violence. It is in these matters where the two countries' experiences critically diverge. In both cases, language of territory, ethnicity, and nationalism – all factors that became valuable political resources because of past and present state initiatives – helped mobilise the attacks. However, the structure and objectives were fundamentally different as will be the lasting impacts on the nature of politics and society. Whereas the Kenyan violence was about controlling the state, the xenophobic attacks were a further reflection of a political culture that is territorialized but decidedly anti-state. This is not a call for anarchy, but the continued efforts to resist the centralization of power in a single institution or person (cf. Götz 1995).
Although some of the Kenyan violence had only questionable links to political structures and there was opportunistic looting, the attacks were driven by central political objectives. In almost all instances, the violence was encouraged or allowed with the goal of unseating Kibaki's government to make way for Odinga's Orange Democratic Movement. Equally importantly, the violence remained more or less controlled by actors who are not firmly entrenched within central state structures. According to the International Crisis Group's initial evaluation (2008), “state authority collapsed in the political strongholds of the opposition Orange Democratic Movement (ODM). Supporters of its leader […] took to the streets in violent protest […]”. Even if the attacks do indicate a loss of authority, it is not the state's centrality that was in question. Rather, what occurred in Kenya is “normal” politics by other means. Had the protesters remained peaceful in their protests against a stolen election, they would have been applauded. While the violence challenges a strict Weberian definition in which the state retains a monopoly of the legitimate use of force, this group was not out of control or seeking to destroy Kenya's state structures. Rather, they were supporting a universally accepted and legitimate political candidate vying for a state position a candidate who was ultimately able to stop the violence. This is not the collapse of a state-centred authority system, but a challenge to the sitting president's authority (see CIPEV 2008: viii).
Whereas the Kenyan violence was both national and state-centric, South Africa's xenophobic violence was decentralised and rooted in the micro-politics of township life. Our research found no evidence that attacks against foreigners were planned and orchestrated by a single organisation or individual across the sites despite early talk of a central “third force” (see Mhlana et al. 2007). Rather, it was organised and led by local groups and individuals as an attempt to appropriate localised state authority for political and economic benefits. There were instances of cooperation between the various sites –“comrades” 1 from Itireleng (Laudium) helped organise attacks in Atteridgeville in March 2008 and there were attempts by groups from affected areas to attack or influence attacks in non-affected areas – but these were local gangsters exploiting anti-outsider discourses: using the legacy of state divisions but without the goal of controlling the state.
Comrades are a self-appointed leadership group in Itireleng.
This fragmentary violence was made possible by the absence of trusted, legitimate central leadership. In its absence, self-appointed structures almost completely appropriated the authority constitutionally mandated to local government structures to operate as an “untouchable” parallel leadership. Much as the presidency in Kenya is presumed to be linked to enormous wealth (CIPEV 2008: viii), community leadership is an attractive alternative for the largely unemployed residents of the informal settlements. It is a form of paid employment or an income-generating activity whereby supposedly voluntary leaders often charge for services, levy protection fees, sell or let land and buildings, and take bribes in exchange for solving problems or influencing tender processes.
The profitability of community leadership positions has attracted considerable infighting and competition for power and legitimacy among different groups present in affected areas. Indeed, street committees, Community Policing Forums (CPFs) and South African National Civics Organisations (SANCO) in most areas report involvement in solving all sorts of problems community members bring to them. In Madelakufa II, for instance, respondents report that the CPF, whose mandate is – according to the local CPF leaders – “exclusively fighting crime”, also involves itself in solving socio-economic and service delivery issues. In Du Noon, the local SANCO branch, which the other local leaders call a “family business” constantly battles the ward council when negotiating development projects with donors.
For local political players, organising attacks on and removing the “unwanted” foreigners from affected communities has proven one of the most successful strategies for earning people's trust while gaining additional legitimacy, clients and revenues. Evidence shows that the xenophobic violence in most affected areas was organised by the above-mentioned parallel structures or by some self-serving members of formal institutions, who capitalised on residents' feelings, fears and negative attitudes towards non-nationals. Their help in “resolving” this bitterly felt problem served to demonstrate a superior efficacy in “crime”-fighting and greater empathy with community concerns, thus consolidating their identity as the only “true” leaders. While the objective of the xenophobic violence in South Africa was not the control over the central state, local leaders sought to appropriate local state authority for localised political and economic interests.
Conclusion
The murders, rape, and thievery we saw in Kenya and South Africa are an explicit component of a political project. There was brigandry and opportunism in each, but those neither framed nor legitimised the attacks among those who launched them. Both draw on resources created by past state projects, ethnic or national suspicions, spatialised understandings of rights and belonging, and political structures designed to control critical economic resources. Each of these struggles has been generated by the intersection of long-standing categorisations, normalised modalities of violence, and immediate self-interest. Due to different political structures, the nature of preexisting tensions and conflicts came to be about something substantially different. In Kenya, the violence remained state-centric and largely state controlled. In South Africa, the violence was decidedly opposed to the central state. To some extent this confirms Wimmer's (1997) assertion that xenophobia or other forms of ethnic conflict emerge when a crisis threatens a nation's social compact. In each case, the accepted order was questioned by both domestic and global circumstances. But while Wimmer's argument that these conflicts will be over which groups are entitled to the state's collective goods, the South African case demonstrates that we must look at more than control over the state. Where space is seen as belonging to a group or subgroup, it is not the state over which people necessarily fight – even if they ostensibly mobilise on the basis of nationality – but for semi-autonomous control over sub-national territory.
The account provided above points to a series of normative and theoretical conclusions. The most obvious relate to the nature of politics in the two countries. As Linz and Stepan (1996) argue, democratic consolidation is most successful where political elites work to develop a broadly inclusive political culture. In this regard, the violence and reactions to it provide both room for fear and optimism. As we argue above, a history of violence – particularly violence relying on emotionally powerful categories of belonging such as race, religion, and ethnicity – is a powerful predictor of future insecurity. Without concerted efforts from political elites to bridge differences and create institutional incentives for collaboration, such faults lines provide ready resources for those seeking political or economic gain. The CIPEV report and widespread public debate in Kenya over the nature of the violence is a first step in this direction. But we must not be too sanguine about its possibility for success: President Moi similarly commissioned reports into earlier political violence but did little to address the root causes. The new leadership also seems inclined to forgive and forget in the interest of short-term political expediency. This may promote elite collaboration but will do little to salve local tensions and propensities.
The South African government's response to the attacks augurs poorly for short-term security and long-term democratic consolidation. While the attacks were initially condemned by almost all political leaders, there are few reasons to think that they will not happen again. Protecting foreigners' rights, dignity, and welfare ranks near the bottom on the country's list of political priorities, far below debates over the future president or the break away political party. Despite calls for unity, few speak of the need to make space for “outsiders” within South African communities. Even fewer speak of the need to ensure that segments of the population are no longer institutionally and socially excluded from basic rights to security and welfare. For these outsiders – foreigners as well as some South Africans – their rights to space and life continue to be subject to the whims of their neighbours and local leaders. Preliminary reports suggest that while elites continue to call for unity, local leaders are recognising the enormous political points to be won by demonising outsiders.
The nature of the violence, and the centrality of local political leaders and structures, points to an even more critical aspect of contemporary South African politics. While drawing on nationalism's power to define and exclude, the driving forces behind the violence were sub-national and, in many instances, work directly against the consolidation of national political or institutional frameworks. At one level, such political and institutional fragmentation all but insures South Africa against national conflict or ethnic cleansing. But while avoiding single, orchestrated incidents, South Africa is likely to witness the continuation of deadly and destructive, if small scale, outbreaks. Indeed, the violence that happened in May has continued in other forms. With the global economic downturn and heightened expectations for the new leadership, such attacks seem almost inevitable.
It is in the continued power of local politics that this article makes its most significant analytical and theoretical contribution. Linz and Stepan's approach, along with those of many others (see Whitaker 2005), continue to speak of politics as a fundamentally national process and project. By allowing themselves to become ensnared in what Agnew (1999) famously terms the territorial trap, such analysts miss the importance of subnational variations not only in the nature of the state (cf. Boone 2003; Herbst 2000) but in the nature of nationalism and ethnicity. While it is true, as Pzeworksi et al. (1995) argue, that weak states will tend to encourage ethnic mobilisation, what happened in South Africa was more than that. The mobilisation we recently witnessed not only had ethnic components, but also simultaneously drew on discourses of nation and the right to space. This is not only political opportunism, or petty criminals working within the cracks of national politics (cf. Mueller 2003). It is instead a fundamentally different form of politics that brings together nation, ethnicity, state, and territory in novel ways that, while creative, are also deeply dangerous.